n8n tutorial for beginners: build your first AI workflow

To build your first AI workflow in n8n, sign up for n8n Cloud or self-host the free Community Edition, create a workflow with a trigger such as a form submission, add an AI node to classify or summarise the input, then route the result to tools like Slack, Google Sheets or your CRM. Test with sample data, add error handling and activate it. n8n Cloud is billed by workflow executions, and the Community Edition is free to self-host.

What is n8n?

n8n is a workflow automation platform with a visual editor. You connect nodes on a canvas: a trigger starts the workflow, and each following node does one job, such as reading a spreadsheet, calling an API, asking an AI model a question or sending a message. It can be used in n8n’s cloud or self-hosted for free. Our comparison of n8n, Zapier and Make explains how it differs from other tools.

How do you get started?

Choose cloud or self-hosted: n8n Cloud is quickest; self-hosting is free but needs a server.
Create a workflow: Open the editor and add your first node.
Add credentials: Connect the apps you need, such as Slack, Google Sheets and an AI provider.
Learn the panel: Each node shows its input on the left and output on the right, which makes debugging easy.

How do you build the lead-triage workflow?

Add a trigger: Use the n8n Form Trigger or a Webhook node connected to your website form.
Add an AI node: Use the AI Agent or a basic language-model node. Ask it to label the lead hot, warm or cold and write a one-line summary, returning JSON.
Parse the result: Use the output parser or a Set node to turn the AI reply into fields.
Route by label: Add an IF or Switch node to branch on the label.
Notify and log: Post hot leads to a Slack channel and append every lead to a Google Sheet.
Add error handling: Set an error workflow that alerts you if a step fails.

The AI node is where n8n differs from simple automations: it can read free text, such as a message field, and make a judgement a rule-based filter could not. Keep the prompt short, give examples of each label and ask for a fixed JSON format.

How do you test safely?

Pin sample data on the trigger so you can rerun without new submissions
Run each node individually and check its output
Test awkward inputs: blank messages, spam, other languages
Point Slack to a test channel until you are confident
Only then activate the workflow

What will it cost?

n8n counts one execution per full workflow run, however many nodes it has, so a lead form with 30 submissions a day is roughly 900 executions a month. Check current plan allowances on the n8n pricing page. AI model usage is billed separately by your AI provider. If you prefer to run it yourself, see how to self-host n8n.

What n8n concepts should beginners know?

Nodes: Each box on the canvas does one job: read, transform, decide or send.
Items: Data flows between nodes as a list of items, one per record.
Expressions: Small formulas that pull a value from an earlier node into a field.
Credentials: Saved, encrypted connections to your apps, reused across workflows.
Executions: The log of each run, with the data at every step.
Error workflows: A separate workflow that runs when another fails.

Most beginner problems come from data shape: a node expects one item and receives many, or a field name changes. Opening the execution log and reading each node’s input and output usually reveals the issue within minutes.

What should you build next?

A daily digest that summarises new support tickets
Invoice data extraction into your accounting tool
Meeting notes summarised and sent to the CRM
Social mentions classified by sentiment

For more ideas, see our 30 AI automation ideas for small businesses, and for the wider picture, what AI automation is.

Once your first workflow has run reliably for a couple of weeks, review its executions, tighten the AI prompt where labels were wrong and add a second workflow. Small, dependable automations build confidence faster than one large, fragile one.

Keep a short note of what each workflow does and who owns it.

How UIUX Studio helps

We build n8n workflows and AI agents for teams, from first automations to self-hosted production setups. Talk to our Agentic AI & Automation team.

Frequently asked questions

Is n8n free?

The self-hosted Community Edition is free; n8n Cloud is a paid plan billed by workflow executions.

Do I need to code?

No for most workflows; small JavaScript snippets help with advanced cases.

What is an execution?

One full run of a workflow, however many nodes it has.

Can n8n use ChatGPT or Claude?

Yes. AI nodes support major model providers.

How long does a first workflow take?

A simple one can be built and tested in an afternoon.